OXIUATION OF LOWERPARAFFIN HYDROCARBONS
Nov. 5 , 1956
will involve hydrogen ions. An analogous situation was encountered by King and Pandowis while studying the oxidation of bromide ion by cerium(1V). The following mechanism is postulated to explain the reduction of chromium(V1). I t is based on the reactive species being HCr04since it is the predominant species in these solutions. The steps are (15) E. L. King and M . L. Pandow,
T I i I S JOIJRNAL.
[CONTRIBUTION FROM
THE
75, 3063 (1953).
+ + + +
H + HCrO4- = H2CrO4 H z C r 0 4 e- = H2Cr0,H+ H2Cr04- = HJCrOI HlCr04 e- = H & r 0 4 Cr(1Y) = Cr(II1)
5553 fast fast fast slow fast
Acknowledgment.-The authors wish to thank E. I. du Pont de Nemours and Co., and the Wisconsin Alumni Research Foundation for financial assistance. MADISOS,WISCONSIN
FRICK CHEMICAL I,ABORATORY,
PRINCETON
UNIVERSITY]
The Oxidation of Lower Paraffin Hydrocarbons. 11. Observations on the Role of Ozone in the Slow Combustion of Isobutanel BY CLARENCE C. SCHUBERT, S.J., AND ROBERTN. PEASE RECEIVEDAPRIL 12, 1956 The reaction of ozonized oxygen (ca. 3-6 mole % 0,) with isobutane in the temperature range 110 to 270' was compared with the slow uncatalyzed reaction with oxygen alone. The ratio of gram atoms of oxygen fixed in liquid product t o moles of ozone added increased from 3.4 at 125" t o 4.2 a t 200" and 5.0 a t 225". The ozone-induced oxidation merges into the normal slow combustion reaction at ca. 265". Approximately one third of the condensed products of the isobutane-ozone reaction a t 150" was found t o be t-butylhydroxymethyl peroxide. The same peroxide was indicated to be a product of the isobutane-oxygen reaction at 270" by a comparison of infrared spectra. It is proposed t h a t ozone may be the active intermediate responsible for chain branching during the slow combustion of hydrocarbons in oxygen. Ozone might result from the reaction: Rot. 0 2 +0 3 RO. Ozone probably has the stability requirements to account for the cool flame reaction and negative temperature coefficient region observed in the combustion of hydrocarbons. Preliminary attempts to detect ozone during the normal slow reaction by observing the ultraviolet absorption in a 3-meter tube were unsuccessful due to general absorption in the 2500 A. region.
+
+
The preceding paper in this series2 described the reaction of ozonized oxygen with paraffins a t low temperatures, ix.,25-50'. This study has been extended to include the temperature range 110270". At the higher temperatures employed the reaction with ozonized oxygen merges into the uiicatalyzed slow combustion reaction. In this paper the oxidation of isobutane in the presence of ozone and in its absence are compared, with attention paid to the identification of peroxidic products. Recently, Batten, et u Z . , ~ have investigated in a static system the slow oxidation of isobutane, and Neu4 has reported on the oxidation of n-butane. Current theories relative to the slow combustion of hydrocarbons favor the postulation of peroxides or of aldehydes as the intermediates responsible for the termination of the initial induction period and for chain branching. These theories are reviewed by Batten, et d 3 It is of interest that these authors conclude that peroxides in their normal state are not the substances responsible for the termination of the first induction period. This work may (1) Taken from a thesis submitted by C . C. Schubert S.J. in partial fulfillment of the requirements for the P h . D . degree. T h e work described in this paper was jointly supported by Contract NOrd-79?0 with the U. S. Naval Bureau of Ordnance a s coijrdinated by t h e Applied Physics Laboratory. T h e Johns IIopkins University; and by Contract NO-ori-105 with t h e Office of h-aval Research a s coardinated by Project Squid, Princeton University. Reproduction, translation, publication, use and disposal in whole or in part by or for t h e United States Government is permitted. (2) C . C. Schubert S.J. and R. N. Pease, THISJOURNAL, 7 8 , 2044 (1950).
(3) J . J . Batten, H. Gardner and M. Ridge, J . Chem. Soc., London, 3029 ( 1 9 5 5 ) ; J. J. Batten and M. Ridge, Aiisfralian J . Chem.. 8 , 370 (19.55). (4) J. T. Neu, Abstracts 127th ACS Meeting, April 1055, paper No.
50, p , IGP.
Also, J . Phyr. Chem., 6 0 , 320 (1956).
aid in deciding whether ozone is a likely candidate for the role of intermediate. Three aspects of the slow oxidation reaction were considered: (1) the formation of liquid products from the reaction of ozonized oxygen with isobutane in the temperature range 110-270"; ( 2 ) the reaction without added ozone a t ca. 270"; (3) the identification of the peroxidic material formed in the isobutane-ozone and the isobutane-oxygen reaction. Experimental Method Isobutane (C.P. grade from Matheson Co.) was mixed with varying amounts of ozonized oxygen, then passed through a heated glass reactor in a flow system. Condensible products were collected in a trap kept at 0" and weighed. The rate of production of this liquid condensate was taken as a rough measure of the extent of reaction. The reactor wasa 5-cm. Pyrex cylinder of 640-ml. volume. Prior to initial use it was cleansed with hot concentrated nitric acid followed by rinses with distilled water and a final rinse with a saturated solution of boric acid. The reactor was mounted in a vertical position in a furnace; the reactant gases without preheating entered at the top and products emerged a t the bottom where they were collected in a trap immersed in ice-water. Temperatures were measured by means of a thermometer placed in contact with the reaction vessel. Due to the exothermic nature of the reaction and the impossibility of preheating the ozone more rigid temperature control was precluded. Commercial tank oxygen was ozonized by means of a Siemens type ozonizer constructed of Pyrex and provided with a grounded condenser jacket filled with circulating cooling water. Ozonized oxygen was mixed with isobutane in a 2 mm. capillary tube, then introduced directly into the reactnr by a 5-cm. length of this same2 mm. capillary. Ozone concentrations were measured by an infrared absorption method calibrated, in turn, by an absolute method similar t o that described by Jahn5 (decomposition a t a heated platinum filament). ( 5 ) S Jahn, Ber., 43, 2319 (1910).
Results 2G5-275' was reached. From this point onward Thc effect of the addition of varying amounts of the reaction went virtually to completion, with or without added ozone (see curve 13, Fig. 1 ) . O / ( J 1 l C to isobututic oxygen mixtures is ~ r i a r l ccle,ir I t should be emphasized tllat the above ratios 1)) :i consideration of the data shown in Ijig. 1. Curves A,B and C of Fig. 1 show the production of are based only on liquid condensed a t 0". For liquid condensate :is a function oi temperature for this reason, analogous ratios obtained in the earlier mixtures containing varying amounts of oxygen work a t 2.7-50" (ratios of 1.3-1.6) are not directly aiid 07oiie (see caption, Fig. 1) Curve D was ob- coinparable. The latter, based on infrared spectained when no ozone was adrlcd to the introduced troscopy, included carbon oxides but not water. oxygen. So reaction is noticeable below 200" with- Se~-crthelcssit is clear that the ozone reaction out o7one. lloreorer, it was noted that the reac- triggers in n limited degree the normal oxidation tion with added ozone, which is already well ad- chain at tenipcraturcs substantially below the nori a i c e d a t 2 Z " , mould gradually cease entirely 11131 range. Experiments similar to those described above upon stopping the discharge in the ozonizer. 113th added ozone the production of liquid condensate were perforined with n-butane and propane. The gradually increases with rise in temperature until reaction between ozonized oxygen and n-butane a t Xi.5-2ij" the reaction merges into the slov osi- proceeds initially slower than in the case with isodation without o7oiic (see curve). O ~ o n cconwiiil)- butane, but tlie slow combustion reaction becomes tion rvas virtually complete abovc 110' for the r('qi- sell-stistaining a t :I lev-er temperature, 255". Prop:iiie behaves in :L comparable inanner and betlmce times einployed, 3 5 to 7 niintites conies self-sustaining a t 2%". -1point worthy of r eniph:isis is that which was noted also by Briner." Ozonc completely elirnin:ites the first induction period. I
[j
JJ-2
I
I
.
1
1
/
1
130 200 2 50 mn Tcnip , "C Production of liquid coiidensate from the reaction of ozonized oxlgcri n itlt ~ s o h ~ i l a n c 100
Fig 1
H c , ml /set
A B C
Mole c ( 0 In
0 7s 4 2
1
1 1 1
D
o
ml /bet
02
3 2 1 9 6 0
TS
0
u
loner estimate (condensed liquid products only) of the efficiency of 07one in effecting oxidation was obtained from the data presented in Fig. 1 , and combustion analyses of the liquid. The latter gave substantially constant coniposition oi H, 39.47, C and 5047; 0 (by difference), mall variations depending on temperature antl on make-up gas The procedure was to coinpare grain-atoms of oxygen in the condensed liquid to gram-moles of 07one reacting The results indicate little vari:ition with gas coniposition at any one temperature Xs tenipcrature was increased, the ratio rose slowly from 3 4 at 1 2 3 O to 4 2 a t 200", and then more rapidly to about .? a t 225°.6 Beyond this temperature the auto catalytic acceleration typical of the reaction nithout ozone set in (see Fig. 1) However, if the o7on11er were cut off, reaction dropped rapidly to negligible proportions until the tcinpcraturc range A\
c nmp 1
~ r tI
Tlrinrr
111
